Automatic generation of prosodic rules for speech synthesis


This paper describes automatic generation of speech synthesis rules which predict the accent component value (stress level) for the bunsetsu in long noun phrases. The rules are inductively inferred from a lot of speech data by rising two kinds of treebased methods. the conventional tree generation and the SBR-tree algorithm. The rule sets automatically generated by two methods have the almost same performance and decrease the prediction error to about 14Hz from 23Hz of the accent component value. The rate of the correct reproduction of the change, that is increase or decrease, for adjacent bunsetsu pairs is also used as a measure of evaluation and the generated rule sets correctly reproduce about 80% of the change. Effectiveness of the rule sets is verified through the listening test. And, the SBR-tree methods generate very compact rules which are easy for human experts to interpret and match with the former studies.

DOI: 10.1109/ICASSP.1994.389224

2 Figures and Tables

Cite this paper

@inproceedings{Yamashita1994AutomaticGO, title={Automatic generation of prosodic rules for speech synthesis}, author={Yoichi Yamashita and Riichiro Mizoguchi}, booktitle={ICASSP}, year={1994} }